Comments and Suggestions

[tmr]

Richard

Nice update.  My guess is that hiding the rating, tags, etc. has balanced any reduction in problem set difficulty.  I haven't done enough problems yet to be sure, but I now realize that I was relying on the problem rating quite a bit before.  I like the bigger board but found the highlighted move squares distracting.  Thanks for making that customizable.  I turned mine off.

Here are some thoughts for some premium and free features that I'd find useful and some comments on some of the features you mentioned in a post yesterday.  First my ideas:

1) I'd think a very attractive premium feature would be the ability to download problems in pgn format.  This way premium users could download them to their handhelds to practice when on the go.  The discontinued problems or I suppose even other rejected problems might be useful even if the don't meet the strict standards for making it into the rated problem set.  From the reduction in your problem set size that amounts to thousands of extra problems.

2) I also think a "daily problem" feature would be attractive where a harder problem is presented (probably with multiple good lines) and users are encouraged to analyze and comment (i.e., direct link between the problem and the discussion forum).  This could be a premium feature but it would probably generate more discussion being a free feature.

3) Another premium feature might be problem solving tournaments where all players within a given rating segment are given the same problem set to solve within a given time limit or for blitz mode in the least amount of time.  Any given tournament could be held over several days or a week and the players could log in and complete their tournament set when convenient.

Some feedback on premium features you mentioned yesterday:

1) I think the premium feature of being able to search the problem set for hard problems would be useful, especially if combined with the ability to download them for analysis away from the computer (i.e., on a handheld on the go).

2) I think an extensive problem history would be most valuable if it was combined with the user's move list so one could go back and see how they solved the same problem in the past.  Without this I don't think I would put much value on an extensive history as I consider the current 1000 problem graph adequate.  Of course I really am only doing standard problems.  Blitz players who do a lot more problems might find a extensive history useful.

3) More statistics is nice but I think need to be combined with some of the above features to make a premium subscription attractive.

4) $20 per year is probably a reasonable price.

[richard]

Thanks for the comments, tmr, I've responded below.( I'm not sure if my previous post referred to this where I go into a few more details:

http://chesstempo.com/chess-forum/news_and_updates/major_site_upgrade-t158.0.html;msg649#new

Or the older post in the thread you contributed to a few days earlier.)

1) I'd think a very attractive premium feature would be the ability to download problems in pgn format.  This way premium users could download them to their handhelds to practice when on the go.  The discontinued problems or I suppose even other rejected problems might be useful even if the don't meet the strict standards for making it into the rated problem set.  From the reduction in your problem set size that amounts to thousands of extra problems.

I'm a bit conflicted when it comes to PGN downloads, I agree it would be incredibly handy to be able to download them (especially in bulk).  It might sound a bit petty but I'm concerned about the ease with which other sites could take all the problems and setup in competition with chesstempo , using my own problem set.  There have been thousands of hours of CPU time spent on problem generation (and re-generation after generator changes), so I'm somewhat protective of the problem set.

The compromise I've arrived at is to supply FEN notation for the start position of each tactic via a "Show start FEN" button on each problem page.  FEN allows users to plug the position into an engine to check their moves after CT marks them wrong. Without the actual move sequence it would be significant work to turn the FEN into a competing problem set (although it would provide a nice starting point, searching a 20-30K of problems for tactics you know are there is much easier than searching millions of games).  The "Show Start FEN" button is already implemented and will be available in the premium rollout as soon as I get a few bugs and a payment system setup.

2) I also think a "daily problem" feature would be attractive where a harder problem is presented (probably with multiple good lines) and users are encouraged to analyze and comment (i.e., direct link between the problem and the discussion forum).  This could be a premium feature but it would probably generate more discussion being a free feature.

This is a great idea.  What I'd really like to do is integrate the rest of the site much more closely into the forum, allowing arbitrary board embedding into forum posts, as well as easily embedding the tactics by problem number.  Once this is done the daily problem feature would be easy to do well.

3) Another premium feature might be problem solving tournaments where all players within a given rating segment are given the same problem set to solve within a given time limit or for blitz mode in the least amount of time.  Any given tournament could be held over several days or a week and the players could log in and complete their tournament set when convenient.

I really like this idea, I've been kicking around a variation of this idea myself for a while now (my idea was more one-on-one rather than a group of users in a tournament , but I think the tournament is a great idea)

Some feedback on premium features you mentioned yesterday:

1) I think the premium feature of being able to search the problem set for hard problems would be useful, especially if combined with the ability to download them for analysis away from the computer (i.e., on a handheld on the go).

2) I think an extensive problem history would be most valuable if it was combined with the user's move list so one could go back and see how they solved the same problem in the past.  Without this I don't think I would put much value on an extensive history as I consider the current 1000 problem graph adequate.  Of course I really am only doing standard problems.  Blitz players who do a lot more problems might find a extensive history useful.

3) More statistics is nice but I think need to be combined with some of the above features to make a premium subscription attractive.

4) $20 per year is probably a reasonable price.

Users' moves are kept after the last update.  One of the upcoming premium features is an ability to view common mistakes made on a per problem basis (and the average rating of the users who made the mistakes). This is not the same as what you've asked for but the data is there for the user to see their own mistakes and for stats to be presented on common mistakes, across multiple types of problems.

The first set of premium features may not be enough to entice users in hoards, but I think it's a reasonable start and I plan to add to the features over time.

Regards,
Richard.

[tmr]

Thanks for the info.  I hadn't considered competitors.  I think making the FEN available is a good alternative for those who want to practice on the go.  The move list is probably secondary since any portable engine should show the best line relatively quickly if needed for checking anyway.

I'm finding the problem set more challanging.  Not sure if its the update or if I'm just coming into a stretch of problems that are more difficult for me.  Seems as if I'm tending to look for the bigger moves and am missing point of the problem.  Maybe this is because you eliminated a bunch of the mating problems and now these type of ones are more prevalent.  This isn't a bad thing since I need to be on the look out for all situations.  This is where the old advantage of having the problem rating showed up.  With a lower rated problem I knew to look for that type of problem.  Now I'm struggling a bit with them.  To me this shows that taking the rating display away was a big improvement for tactics training.  It makes it more frustrating of course but definitely more real life.

I look forward to see the premium feature set and your plans for its role out.

Keep up the good work.

[richard]

Hi Tmr,

For now FEN is only available from each problem view, not downloadable as a set, so it's more useful for users who don't understand a tactic (or believe the tactic has an issue) and want to have their chess engine have a look at the position. 

In terms of perceiving the problem set as more challenging I think it is probably more about the lack of ratings display than anything else. The average rating for problems in the new set is 100-200 points lower than the old set, so that suggests the problems themselves should on average be a bit easier.  There may also be a period of "re-training" your expectations for the new problem set mix. Another factor might be the aprox 2000 new problems that were introduced.  I now have a much better guess at what rating problems should start at by doing a regression analysis on problem length versus rating, however this is not perfect so there may be some new problems with very hard to see solutions but shorter move sequences that are giving you a hard time.

Regards,
Richard.

[slacker00]

Thanks for the info, richard.

I hadn't considered your competitors either.  I guess that's the luxery of being a humble user. lol

An idea might be to  only allow one problem per day to be exported, or maybe only  one problem export for every 100 attempts.  Those kind of limitations could thwart any kind of automated problem set theft.  I think  I could live with that kind of limitation.  If anything,  it could motivate me to solve more  problems, because I'll know after X more attempts, I can export that one that's been nagging me for further analysis! haha

[richard]

For analysis purposes the FEN available for each position should be enough. I think it's only when you want to start playing through the tactics offline and need to know the "answer" that PGN becomes important.

Richard.

[tmr]

I agree that for analysis purposes the FEN should be enough.  An alternative might be to allow the player to click a button that allows analysis right on the problem board.  Since you've already have the computer option that shows the best line continuation it seems not much of a stretch to add an option to allow a player to click "analyze this problem" and have the computer respond to the "wrong" moves.  Of course no ratings for this.

I'd still like see some premium option for bulk FEN download.  Slacker00 has a good idea tying the number of downloads to the participation on the site, though a small number, say one a month, even for full PGNs wouldn't really have much additional value given that the FEN was already available.  I'd use this type of option for practicing problems when commuting and at lunch so to be of value this would probably mean allowing downloads of about a couple hundred a month.  As I mentioned before, using the rejected problems could be an alternative here.  As it appears that there are thousands of those parceling them out at a couple hundred a month could be an attractive alternative for premium membership for those seeking this option.

Of course I know one site can't be everything and time is also a factor.  I suppose with a bit of looking I can find something similar out there already.

[richard]

Hi Tmr,

The "analyze this problem" option is something I've considered before, at the moment I don't really have the computing resources to make this feasible. I could have a small chess engine that runs on the browser but Javascript chess engines are very weak , probably too weak for this purpose. Java is another option, but could get a little messy. Probably not the next version of the generator, but the one after will save not just the best line but the alternative lines and allow the "computer analysis" button to show those as well which provides a little more detail (but doesn't address why the user made a bad move when it wasn't in the top N lines).

Bundling PGNs is certainly an option, but there are still a few other items ahead of it on the todo list for now.  In terms of commuting etc I think it would be cool if there was a version of chess tempo that was able to be run on mobile devices and still connected to the CT server so rated problem solving could be done on the run. Of course not all countries have ubiquitous mobile access at rates that would make this an attractive option.

Regards,
Richard.

[rockyrook]

I simply wanted to express my gratitude to you for providing this site!  I've really enjoyed working on tactics here.

I think the improvements are awesome!  I was thrown off a bit when I first saw that the tags and avg time solved did not appear while I was working on the problem.  But I've gotten used to it and I like it.  I'm not distracted any more by looking for comments and tags and time while trying to solve the problem.  It makes me more honest!

I was searching the forum for information about having to pay to use premium features.  I'm a bit excited about this because now you'll get some compensation for your hard work and because we get more neat features!  In particular, I like the idea of the FEN being available.  When I'm solving standard problems, I almost always copy the tactics I get wrong into a database I've set up, so I can go back and work on those problems.

And one other comment ... I'm not sure if there is anything you can do about this, but I first stumbled on Chess Tempo when I was looking for an alternative to CTS because CTS is blocked by Websense at my work.  I found Chess Tempo almost a year ago and Websense has yet to block it.  I don't know how Websense works and like I said, I'm not sure there's anything you can do about it, but if there is, please try to keep Chess Tempo "under the radar" if you will.

Thanks!

-RR

[richard]

Thanks rockyrook,

Nice to hear you are enjoying the recent changes.

The premium features are still undergoing testing.  I'd also like to have a new problem set update done before the final release of the premium features, but I might not be patient enough to wait for that :-)

In terms of the database of incorrect problems, one of the premium features involves "dynamic problem sets" which provides automatically updated problem sets based on a set of search criteria.  So far example you can have a problem set that only includes the problems you have got wrong or only the problems you keep getting wrong. As you solve these problems the problem set gets smaller and as you get problems wrong in standard/blitz the dynamic "problems I keep getting wrong" set will automatically have the new mistakes added.

The "problems I keep getting wrong" set is one of a new set of default dynamic problem sets, but you can also create your own variations, so for example you could create a new  dynamic problem set which holds all the problems you have got wrong that are rated below 1500 and have been tagged as "forks".

Hopefully Chess Tempo manages to stay out of the websense blocking list.  I imagine the CTS blocking was a mistake although maybe they saw CTS as something that could impact worker productivity during work hours so banned it :-)   

Regards,
Richard.